Efficacy of Colchicine in the Treatment of COVID-19 Patients: A Systematic Review and Meta-Analysis

Objective: We assessed the efficacy of colchicine in COVID-19 patients through a systematic review. Methods: Six databases were searched until March 2022 for studies assessing colchicine versus control in hospitalized patients with COVID-19. The primary outcome was mortality, and secondary outcome was length of hospitalization. Inverse variance and random effect meta-analyses were performed. The strength of evidence was assessed using GRADE. Results: Nine studies (five randomized clinical trials (RCTs) and four non-randomized studies of intervention (NRSI); n = 13,478). Colchicine did not reduce mortality in comparison with the standard of care in RCTs (RR 0.99; 95%CI 0.90 to 1.10; p = 0.90); however, it did reduce mortality in NRSI studies (RR 0.45; 95%CI 0.26 to 0.77; p = 0.02). In the analysis of RCTs, colchicine did not reduce the length of hospitalization in comparison with the standard of care (MD: −2.25 days; 95%CI: −9.34 to 4.84; p = 0.15). Most studies were scored as having a high risk of bias. Quality of evidence was very low for primary and secondary outcomes. Conclusion: Colchicine did not reduce the mortality and length of hospitalization in comparison with the standard of care in hospitalized patients with COVID-19. The published evidence is insufficient and of very low quality to recommend treatment in patients with COVID-19.


Introduction
The worldwide pandemic caused by COVID-19 keeps the scientific-medical community uncertain due to the lack of a specific treatment protocol because there is no cure for the coronavirus [1]. In our country, Peru, treatment with colchicine has not been established for outpatients or hospitalized patients with COVID-19. Thus, in the last year, experimental vaccines and treatments have been studied and developed to combat SARS-CoV-2, as well as the characteristics of the viral infection [2]. Similarly, among the treatments evaluated in published clinical trials, immunomodulatory drugs against inflammatory reaction and cytokine storm in patients with severe and critical infection stand out. [3].
Colchicine, an immunomodulatory drug that acts by inhibiting microtubules, is widely used in conditions such as gout and those involving local tissue inflammation [4]. The rationale for using colchicine in patients with COVID-19 is based on the activation of the NLRP3 inflammasome by viroporin E, a component of SARS-CoV-2 that generates an inflammatory response [5]. Since colchicine inhibits the NLRP3 inflammasome, it has been postulated for use in SARS-CoV-2 infection [6].
The objective of this study was to assess the efficacy of colchicine in COVID-19 patients through a systematic review and meta-analysis.

Materials and Methods
This systematic review was reported following a PRISMA 2020 statement [7]. We assessed studies that evaluated the efficacy of treatment with colchicine in patients with COVID-19. The protocol was registered in the PROSPERO database (CRD42021230362).

Data Sources
We searched in PubMed, Scopus, Web of Science, Ovid-Medline, Embase, and Cochrane Central Register of Controlled Trials (CENTRAL) until 10 March 2022. The search strategy for PubMed was adapted for use in the other databases (Table S1). There were no restrictions on language or publication date. We hand-searched reference lists of all included studies and relevant review articles to identify other potentially eligible trials. Additionally, we searched in the trial registries ClinicalTrials.gov (https://www.clinicaltrials.gov/, accessed on 10 March 2022), WHO International Clinical Trials Registry Platform (http://apps.who. int/trialsearch/, accessed on 10 March 2022), and a preprints/preproofs repository for finished as well as ongoing trials.

Eligibility Criteria
Studies were included if they met the following criteria: (i) randomized clinical trials (RCTs) and observational studies (non-randomized studies of intervention (NRSI)); (ii) hospitalized adult patients (≥18 years of age) diagnosed with COVID-19 as defined according to study authors; (iii) any dose and duration of colchicine as experimental/intervention group; (iv) placebo or standard of care as the control group or comparator. We excluded studies according to the following criteria: systematic reviews, narrative reviews, conference proceedings, editorials, case reports, case series, letters to the editor and abstracts.

Study Selection
One author (J.J.B.) downloaded all titles and abstracts retrieved by electronic searching to the Rayyan platform and duplicate records were removed. Titles and abstracts were independently screened for relevance by three review authors (J.B.M., D.B.P. and C.S.M.G.) and any disagreements were resolved by a fourth review author (J.J.B.). We retrieved the full text of selected trials and three authors (J.B.M., D.B.P. and C.S.M.G.) independently screened the full text, identified studies for inclusion, and registered reasons for the exclusion of studies. We resolved any disagreement through consulting a fourth review author (J.J.B.).

Outcomes
The primary outcome was mortality, and the secondary outcome was length of hospitalization.

Data Extraction
Three review authors (J.B.M., D.B.P. and C.M.) independently carried out data extraction using a data extraction form that was previously piloted on at least one study in the review and any disagreements were resolved by a fourth review author (J.J.B.). If additional data were needed, we contacted the corresponding author through email to request further information. We extracted the following study data from full-text articles: first author name, year of publication, study design, study location, study design, eligibility criteria, sample size, age, sex, description of intervention and control groups, primary and secondary outcomes.

Risk of Bias Assessment
Two investigators (D.B.P., C.S.M.G.) independently assessed risk of bias (RoB) by using the Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) tool for NRSI [8] and the Cochrane Risk of Bias 2.0 tool for RCTs [9]; disagreements were resolved by discussion with a third investigator (J.J.B.). RoB per domain and study was described as low, moderate, serious, critical and no information for cohort studies, and as low, some concerns, and high for RCTs.

Statistical Analysis
Inverse variance and random effect meta-analyses were performed to evaluate the effect of colchicine vs. control on outcomes when outcome data were available for at least two RCTs or NRSI judged to have homogeneous study characteristics. Effects of meta-analyses were reported as relative risks (RR) and their 95% confidence intervals (CIs). CIs of effects were adjusted with the Hartung-Knapp method [10], and the between study variance tau2 was calculated with the Paule-Mandel method [11]. The effects of colchicine were described with log relative risks (LogRRs) with 95% confidence intervals (Log RR 95% CIs) for dichotomous outcomes in the NRSI studies that were evaluated. The RR (TE) and standard error (seTE) were calculated for the effect value in each study. Heterogeneity of effects among studies was quantified with the I 2 statistic (an I 2 > 60% means high heterogeneity). In sensitivity analyses, we assessed (i) all meta-analyses performed without the Hartung-Knapp adjustment and (ii) only studies with a low risk of bias. The R 3.5.1 meta-package was used for all meta-analyses. Statistical significance was set with a p-value < 0.05.

GRADE Quality of Evidence
The quality/certainty of evidence was evaluated using the GRADE methodology, which covers 5 aspects: risk of bias, inconsistency, indirectness, imprecision, and publication bias [12]. Quality of evidence was evaluated per outcome and described in Summary of Findings (SoF) tables; GRADEpro GDT was used to create SoF tables [13].

Ethical Considerations
This is a systematic review of published and open information in which no human subjects participated. Thus, no ethics committee approval was required.

Risk of Bias Assessment
Overall, two RCTs were scored as high RoB [18,19] and one was scored as some concerns RoB [20]. One study had a high RoB in deviations from the intended interventions [18], and one study had a high RoB in missing outcome data ( Figure S1). In NRSIs, two studies were scored as serious ROBINS-I [14,15] and one was scored as critical ROBINS-I [17].

Effect of Colchicine in Primary Outcomes
In the analysis of randomized controlled trials, colchicine did not reduce the mortality in comparison with the standard of care (RR 0.99; 95%CI 0.90 to 1.10; p = 0.90; I 2 = 0%, Figure 2a). However, in the analysis of NRSI studies, colchicine reduced mortality in comparison with the standard of care (RR: 0.45; 95%CI: 0.26 to 0.77; p = 0.02; I 2 = 22.6%; Figure 2b).

Risk of Bias Assessment
Overall, two RCTs were scored as high RoB [18,19] and one was scored as some concerns RoB [20]. One study had a high RoB in deviations from the intended interventions [18], and one study had a high RoB in missing outcome data ( Figure S1). In NRSIs, two studies were scored as serious ROBINS-I [14,15] and one was scored as critical ROBINS-I [17].

Effect of Colchicine in Primary Outcomes
In the analysis of randomized controlled trials, colchicine did not reduce the mortality in comparison with the standard of care (RR 0.99; 95%CI 0.90 to 1.10; p = 0.90; I 2 = 0%, Figure  2a). However, in the analysis of NRSI studies, colchicine reduced mortality in comparison with the standard of care (RR: 0.45; 95%CI: 0.26 to 0.77; p = 0.02; I 2 = 22.6%; Figure 2b).

Effect of Colchicine in Secondary Outcomes
In the analysis of RCTs, colchicine did not reduce the length of hospitalization in comparison with the standard of care (MD: −2.25 days; 95%CI: −9.34 to 4.84; p = 0.15; I 2 = 0%; Figure 3). For the other prespecified secondary outcomes in the protocol, there was insufficient information among the included studies to analyze the effects of colchicine on clinical improvement, the need for mechanical ventilation, transfer to the intensive care unit, serum levels of inflammatory markers, (C-reactive protein (CRP), D-dimer (DD), ferritin and lactate dehydrogenase (LDH)), serum levels of cardiac markers (troponin), and adverse effects.

Effect of Colchicine in Secondary Outcomes
In the analysis of RCTs, colchicine did not reduce the length of hospitalization in comparison with the standard of care (MD: −2.25 days; 95%CI: −9.34 to 4.84; p = 0.15; I 2 = 0%; Figure 3). For the other prespecified secondary outcomes in the protocol, there was insufficient information among the included studies to analyze the effects of colchicine on clinical improvement, the need for mechanical ventilation, transfer to the intensive care unit, serum levels of inflammatory markers, (C-reactive protein (CRP), D-dimer (DD), ferritin and lactate dehydrogenase (LDH)), serum levels of cardiac markers (troponin), and adverse effects.

Sensitivity Analysis
Sensitivity analysis showed no differences with the primary analysis for the outcomes evaluated ( Figure S2).

Quality of Evidence (QoE)
QoE was very low for primary and secondary outcomes (Table S3). In mortality, for

Sensitivity Analysis
Sensitivity analysis showed no differences with the primary analysis for the outcomes evaluated ( Figure S2).

Quality of Evidence (QoE)
QoE was very low for primary and secondary outcomes (Table S3). In mortality, for RCTs and NRSI studies, the QoE was very low due to high risk of bias, the heterogeneity among the studies, and the imprecision of the effect. In length of hospitalization, the QoE was very low due to the high risk of bias and the imprecision of the effect.

Main Findings
In our systematic review, we found that, in RCTs, colchicine did not reduce the mortality and length of hospitalization in comparison with the standard of care in hospitalized patients with moderate to severe COVID-19. However, in NRSI, colchicine reduced mortality in comparison with the standard of care. Most studies were scored as having a high risk of bias and the quality of evidence was very low for all outcomes.
Colchicine is a microtubule inhibitor that has been proposed as a possible treatment for COVID-19 patients based on the following mechanisms: (a) changes in SARS-CoV-2 viral replication due to changes in microtubules, which are important for intracellular essential transport and the creation of double-membrane vesicles [5]; (b) inhibition of the NLRP3 inflammasome with a decrease in interleukin (IL)-1β and the consequent reduction of several pro-inflammatory cytokines that are produced in excess in COVID-19 patients [23]; (c) reduced expression of L-selectin inhibits neutrophil activation, motility, and activation to cell endothelium [24]; (d) inhibition of neutrophil extracellular traps (NET) generation; (e) avoid more leukocyte migration and microvascular thrombosis due to inhibition of complement activation via microtubule disruption, inhibition of C5a, and reduced expression of C5aR; (f) endothelial damage is prevented indirectly by reducing Eselectin-mediated neutrophil adhesion to pro-inflammatory, cytokine-activated endothelial cells and limiting excessive inflammatory activation [25].
Many observational and experimental studies have been conducted based on this to examine the efficacy and safety of colchicine in a variety of COVID-19 scenarios. To examine the evidence from these, systematic reviews and meta-analysis are required. However, owing to the differences in technique, the results of numerous systematic reviews differ from ours.
Salah et al. published a meta-analysis about colchicine in COVID-19 patients and their primary outcomes were all-cause mortality and mechanical ventilation. The study included eight studies (three RCT, five observational, n = 5259). Colchicine showed a reduction in all-cause mortality (RR: 0.62; 95%CI: 0.48-0.8), but it did not reduce the risk of mechanical ventilation (RR: 0.75; 95%CI: 0.45-1.25) [26]. This study included both designs of studies (RCT and observational) in the meta-analysis, therefore the analysis had some concerns. In practice, it is not feasible to include data from different study designs in the same meta-analysis, because this may generate bias or results that are far from the true effect. In our study, the mortality was analyzed by type of study, and we report two effect measures. Another observation is related to the intervention. Our study included only colchicine as intervention; however, in Salah et al.'s work, the authors included studies that analyzed colchicine plus other treatments as an intervention. This may cause bias in the primary analysis of the outcome.
Moreover, Golpour et al. carried out a meta-analysis that included 10 studies (of which four were RCT, n = 5901) to evaluate the efficacy of colchicine in COVID-19 patients. Colchicine was associated with a decreased mortality rate in COVID-19 patients (RR: 0.365; 95%CI: 0.555-0.748) and a decrease in hospitalization time in COVID-19 patients [27]. Both Salah et al. and Golpour et al. did not report the quality of evidence. It is possible that, in both studies, the measure of the effect of "mortality" has been estimated or reported erroneously, therefore it is not possible to have adequate certainty to recommend treatment. Regarding the mortality in both studies, the RCTs included in these studies did not have a significant effect. As with our study, colchicine did not reduce the mortality compared with control based on RCTs effect measures.
Other studies have similar biases in their analysis. For example, Elshafei et al. evaluated nine studies (three RCT, six observational studies, n = 5522) and found lower mortality with colchicine use (OR: 0.35; 95%CI: 0.25-0.48). The qualities of the most included studies were rated as moderate by the authors [28]. Beran et al. analyzed eight studies (three RCT, five observational studies, n = 926) and showed a lower risk of mortality with the colchicine treatment (RR:0.49; 95%CI: 0.34-0.72). The included studies were rated as high quality by the authors [29]. In all previous studies, the authors combined both randomized and non-randomized studies in their meta-analyses. The inclusion of non-randomized studies should be limited to particular situations because the potential biases are greater for non-randomized studies compared with randomized trials when evaluating the effects of interventions [30].
Nawangsih et al. published a systematic review with a meta-analysis to evaluate the effect of colchicine administration on mortality in patients with COVID-19, which included eight studies (three RCT, five observational studies, n = 5530). As with our study, the authors carried out the analysis by type of study. The pooled analysis for observational studies showed mortality reduction (OR: 0.48; 95%CI: 0.28-0.82). However, pooled RCTs did not show this reduction (OR: 0.43; 95%CI: 0.17-1.08) [31]. Unlike the Nawangsih et al. study, our study did not include the COLCORONA trial (n = 4159), as we did not include outpatient studies. Nevertheless, our study included the RECOVERY trial (n = 11,340). This difference in sample size allows our study to have a more precise estimate of the effect of the intervention.
The Pan American Health Organization (PAHO) published a living systematic review and meta-analyses to evaluate the evidence of potential therapeutics options for COVID-19, including colchicine [32]. In this regard, they analyzed five RCT (n = 16,105), including the COLCORONA trial and RECOVERY trial [33]. As with our result in the pooled analysis of RCT, they found that colchicine did not reduce mortality (RR: 1; 95%CI: 0.93-1.08; moderate certainty). In addition, the living systematic review by the PAHO showed that colchicine does not reduce mechanical ventilation requirements (RR: 1.02; 95%CI: 0.95-1.13; moderate certain), though probably reduces hospitalizations in patients with the recently onset disease (RR: 0.8; 95%CI: 0.62-1.03), although the certainly of the evidence was low for this outcome because of imprecision. A living systematic review allows for the updating of the evidence regularly, which is important in the current pandemic scenario.
In our meta-analysis, differences were found between the incidence of sex-matched outcomes. However, other reports, both systematic reviews and other study designs, have not further analyzed the gender-matched analysis, therefore they have not noted a great importance in these differences in treatment with colchicine.
Our study has several strengths. First, we performed a recent and broad systematic search in six databases, two trials registers, and one preprints/preproofs repository without language restriction. Second, we evaluated RCTs and NRSI studies separately; the combination of all types of designs (as occurred in several previous studies) can increase the bias and confusion of the findings. Third, we evaluated the risk of bias from included studies and informed the results together with the certainty of the evidence using the GRADE methodology; meeting the methodological guidance for a high-quality systematic review and meta-analysis is mandatory to increase confidence in the findings. Fourth, we only evaluated studies with hospitalized patients to analyze a population of patients as homogeneous as possible; our findings do not support the use of colchicine in moderate to severe COVID-19. Finally, we also performed sensitivity analysis; the effect was the same as the primary analysis on mortality.

Limitations
Our study also has some limitations. The quality of evidence was very low for the outcomes; the quality of evidence limits the confidence that the estimates of the effect are correct. However, we evaluated the best currently available evidence about colchicine use in hospitalized patients with COVID-19. The result of a pooled analysis from NRSI studies differed from that obtained in the pooled RCTs; this discrepancy may be explained by a higher risk of bias in the NRSI studies that may overestimate the result [34]. Some outcomes were scarce in evaluating their impact in patients treated with colchicine. Well-designed, placebo-controlled clinical trials are needed to increase the certainty of the evidence about the effect of colchicine in COVID-19 patients.

Conclusions
Colchicine did not reduce the mortality and length of hospitalization in comparison with the standard of care in hospitalized patients with COVID-19. The published evidence is insufficient and of very low quality to recommend treatment in patients with COVID-19.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jcm11092615/s1. Figure S1: Risk of bias assessment of included trials. Figure S2: Sensitivity analyses of the effects of colchicine vs standard of care on mortality. Table S1: Search strategy for each engine. Table S2: Ongoing randomized controlled trials. Table S3: GRADE summary of findings table.